Welcome to Netrider ... Connecting Riders!

Interested in talking motorbikes with a terrific community of riders?
Signup (it's quick and free) to join the discussions and access the full suite of tools and information that Netrider has to offer.

[US]Motorcycle ABS, no statistically significant reduction in fatalities.

Discussion in 'Research, Studies, and Data' at netrider.net.au started by robsalvv, Apr 23, 2012.

  1. The motorcycle ABS debate continues... For a summary of two key studies on the topic have a gander at this page: http://thesafetyrecord.safetyresearch.net/2010/11/18/anti-lock-brake-system-debate/

    IIHS study authored by Teoh, argues that there's a 37% reduction in fatalities with bikes using ABS. It and studies like it are used to tout the fatality reduction benefits of ABS. These studies are almost universally flawed and confounded.

    The NHTSA however, using a case controlled methodology instead found a statistically insignificant difference between fatal crashes of ABS and non ABS'd bikes. I can't seem to find the original NHTSA study, but I've found a short summary presented at a recent ENECE conference. http://www.unece.org/fileadmin/DAM/trans/doc/2012/wp29grrf/GRRF-72-25e.pdf

    The conclusion of the case controlled study says:

    To be fair, there was a slight favouring of the ABS'd bikes - but not a statistically significant difference.

    Also, based on what little info there is, the NHTSA's study isn't without it's flaws, one of them being identical to the IIHS study where no fault has been attributed - only the fatal stats were analysed... but I have to say that the conclusion echoes one of my central contentions, ABS isn't a panacea. Get it for the right reasons. Get it with your eyes open.

    On a decent bike it's a $1000 option. A $1000 buys a lot of training that could inherently reduce your risk factors more than ABS.

    Think about it.
    • Like Like x 2
  2. I did think and there's a big confounder here:

    We used the motorcycle models and years identified in the IIHS report as having ABS as an option

    So therefore some of the bikes in the ABS crash group might not have had ABS at all (as it wasn't put on the bike and being expensive probably wouldn't have been on some versions especially the Harleys I would imagine.

    So slightly unfair comparison.

    Might have to chase up the original paper.

    Cheers Spocky
    • Like Like x 1
  3. From the first page you linked to, I think the insurance figures make an interesting comparison because they directly compare the same model of bike with and without ABS.

    I think that all of the studies are going to have difficulty with the self-selecting nature of those who buy ABS bikes. The same thing goes for high-vis vests and white helmets. If you are placing a priority on safety or visibility, then you may ride differently to somebody who chose the open face helmet and rides in a leather vest with no gloves.

    I can't actually see a way around this without giving a thousand riders of similar experience levels the same bike, half with ABS and half without at random, and getting them to ride in the same kinds of conditions for 12 months.

    Sadly, $1,000 doesn't buy all that much training. I'm looking at the Stay Upright Advanced 1 (which I've done before) next month and it is $380.

    (And I actually had a bit of a tumble on the weekend that could probably have been avoided if I had either done a refresher course more recently or had ABS... or better yet, both!)
  4. How do you figure that some bikes might not have had ABS?

    The IIHS study confirmed ABS option via VIN number. So if the NHTSA used the same data set as the IIHS but used a more rigorous approach, wouldn't that make the NHTSA study more valid?

    By the way, the study doesn't include any identifiable sports bikes. Another confounding factor - one of the strongest that all these ABS studies share.
  5. In theory true BUT they don't use the same data set. the NHTSA study uses 2001-2008 data and the IIHS study uses 2003-2008 data. There were 274 non-abs and 47 ABS crashes in the IIHS paper vs 304 and 54 in the NHTSA paper. The IIHS confirmed that ABS was fitted as part of decoding the VIN of that specific bike and thats not so clear from the NHTSA Powerpoint (would like to see original paper that this comes from.

    Was it an option prior to 2008?

    Cheers Spocky
  6. So the NHTSA used a larger data set then... compared them against a control group and determined the influence of ABS. Sounds like you're making a good case for the confidence one could have in the NHTSA study.

    The IIHS instead just directly compared fatal stats of ABS vs non ABS bikes and made a bold assertion of correlation regarding ABS. (Interestingly in the first version of the paper, they had a section talking to the confounds, but that section was dropped in later issues.)

    The power point presentation isn't explicitly clear about how it used the IIHS data - so I agree, getting the original paper would be good - however it doesn't come up on a search of the NHTSA site though. What the power point pdf does say though is "We used the motorcycle model and years identified in the IIHS report as having ABS as an option." So what that tells me is that for 2003 - 2008, the data sets are the same. Then for 2001 and 2002, they only grabbed data for those fatally involved bikes of the same models as in the 2003 - 2008 data, so I think you're assertion about there being additional non ABS option bikes in the data set is wrong.

    I think the intent is very clear even if not explicitly stated.

    Fair point, ABS might have only just started to penetrate sports bikes around this time. I think the fireblade had ABS in 2008, maybe the BMW too. The point to be made though is that all these studies do not feature (and possibly could not feature) sports bikes, yet the studies draw conclusions for all PTW's. This is generally never made clear.


    The vehicles identified include:

    2008 HarleyDavidson VRod,
    20012008 Honda Gold Wing 1800,
    - 20022008 Honda Interceptor 800,
    20012007 Honda Reflex,
    20032008 Honda STl300,
    20032008 Honda Silver Wing,
    20082009 Kawasaki Concours 14,
    20072008 Suzuki Bandit 1250,
    20062008 Suzuki Burgman 650,
    20072008 Suzuki SV650,
    20072008 Suzuki VStrom 650,
    20062008 Triumph Sprint ST, and
    20042005 Yamaha FJR1300
  7. Something which I've not seen mentioned officially wrt ABS but which I can see having a significant effect on the findings is the disconnect between the type of bikes that have hitherto been equipped with ABS and the circumstances where ABS has, IMHO, the most benefit.

    At least until recently, ABS has been the province of upmarket tourers and sports tourers, followed, more recently, by sporty nakeds and full blown sports bikes. Whilst I haven't any figures to back it up, intuitively it seems likely that a large proportion of these bikes will be used in good weather, in open road conditions, by reasonably switched on riders. Circumstances where even a fairly basic level of rider competence will allow braking without dangerous lock-ups.

    Contrast this with the circumstances where ABS will beat the average rider 99 times out of 100. That is greasy, slippery, wet roads, darkness and dense traffic where the unexpected can happen very quickly with limited escape options, cold, tired, wet, distracted riders who just want to get home. IOW, the typical winter (or substitute the appropriate season for the geographical location) evening commute for many. But most of us don't commute on big sports tourers (though some do, and I've done it myself). We're on budget stuff like Ninja 250s, GS500s or middleweight dualsports like my DR, most of which aren't, or weren't until recently, available with ABS.

    I would imagine that it's quite hard design a meaningful study when the market most likely to benefit from ABS inherently has probably the least penetration by the technology, whilst most of the bikes actually fitted with it will never operate in circumstances where there is a significant likelihood that they will ever need it.
    • Like Like x 1
  8. Larger does not mean better. Not the NHTSA also uses an earlier set of data which may have implications of improperly implemented ABS and as it was relatively new technology at the beginning the bikes listed may not have been equipped for a variety of reasons. You may be right BUT the statement hasn't been made therefore you have to make the assumption that the actual bikes checked may not have had ABS but were put into the ABS group.

    They did stretch the interpretation a bit but there did look like a significant difference that was also noted in another paper. If they used the same sets of data then why the different results? Something is fishy.

    They didn't use the IIHS data (they have a bigger set) so you can't make that assumption and as stated they did not in the Powerpoint make the issue that they checked the individual bikes for ABS. Thats a strange thing to leave out when the IIHS paper defined what they did to ensure that the bikes listed individually DID have ABS.

    Same data set (allegedly) and different results? No they have to account for the difference in the later presentation and they don't.

    Agree....they can only compare apples to apples and the sports bike market is very different from the tourer market and they do make the point that the
    number of miles travelled per accident is not listed anywhere.

    Cheers Spocky
  9. LAMS bike riders with and without ABS would be the study to look at as they would be the most likely to slap on the anchors in an uncontrolled manner in a low CC capacity bike.

    The numbers of accidents (fatal) in Australia is very small however in this group now due to the licencing rules/training so it would take a long time to get meaningful numbers.

    Cheers Spocky
  10. Spocky, I'm not a fan of the NHTSA but you are making a specious argument for the sake of it mate. Especially the point that NHTSA could have used differently implemented ABS'd bikes. Like WTF? A bike model is either identified by it's VIN to have ABS or not. There's no grey area. There is no analysis of the type of ABS. Talk about clutching at straws.

    And one of the key points, is that a more appropriate less confounded method produced a result that showed up the confounds of the IIHS approach.

    You cannot defend the Teoh paper surely? I have to ask, do you find the conclusion confronting?

    - - -
    Tapatalking loud, saying somethin'
  11. Even if there is a self selecting effect, most bikes are bought second hand hence those bikes and subsequent crashes may not be from the "cautious" group.

    You wont need to pay $1000 for abs on a second hand bike, most options get no value on the secondhand market and abs would be no different.
  12. I have to apologise as there is actually a point that identifies that they used VIN of actual ABS as opposed to the earlier statement of had ABS as an option.

    We used the motorcycle models and years identified in the IIHS report as having ABS as an option does not specify individual bikes but

    The 2001‐2008 FARS data include 356 motorcycles with optional ABS that was identifiable from the VIN does.

    The issue I am more confused about is how they got such different results from the same dataset.....anyway mea culpa.

    Why would I find it confronting? I just find it interesting that there are two studies using the same data that have come to opposite conclusions and am looking for the reason why.

    On top of the Teoh paper there is also a report by the insurance companies (Highway Loss Data Institute. 2009. Insurance special report: motorcycle antilock braking system (ABS). Highway Loss Data Institute. Arlington, VA.) that showed a 22% reduction in accidents in bikes with ABS vs those without. Thats reported in the IIHS website (so may be bias selection) but insurance companies tend to have their own monetary interests at heart and that also interests me (trying to find the full paper).

    To be honest I'd like to read the full paper that the Powerpoint is based on to see their methodology vs the IIHS study (if you can find it it could you send me a copy, I am interested in stuff like this.)

    So not being specious or argumentative, just wondering the facts behind the two papers and looking for reasons for the differences.

    Cheers Spocky
  13. Fark me, I totally missed this post from you Zenali. :) Welcome to the thread bro.

    Bugger about the spill mate. Hope you and your bike are ok and it's nothing too serious!

    Your point about training made me realise I need to clarify something. The intent of my comment wasn't that any old training would achieve a result, but that properly selected training would.

    If ABS avoids a fall from a braking error and the training in lieu of ABS doesn't sharpen your braking skill, then the training probably wouldn't reduce your risks of a fall from a braking error. ABS would probably be the better choice to address overbraking errors!

    But if you chose training that improved your overall strategies and hazard perception on the road, then that could make a holistic difference and reduce your inherent overall risk of an incident, probably more than ABS would.

    No sweat Spocky. That "option" comment is ambiguous - I read it in the way that it indicated that the "option" had been fitted rather than it having been an available option for a model of bike. I don't think the NHTSA made a rookie mistake and botched up their ABS study by leaving the ABS bit completely a loose end and analysing bikes that had ABS as an option. I think it's pretty clear from the pdf that they did not do that.

    I asked whether the NHTSA conclusion was confronting - because you said you'd thought about ABS, but I bet you didn't think it wouldn't make a difference? It was a stab in the dark.

    Fair question about why the basically same data set produced a different result - in a nutshell, NHTSA used a very different methodology. It's worth understanding that. I'm interested to know whether there were any pitfalls too. The IIHS pitfall is obvious.

    The IIHS and the HLDI have ballsed themselves up. I don't think they're as dumb as they're making out - I'm sure they understand the pitfalls of their studies - but it makes good political mileage for them to keep running with their lines. HLDI are the dumber though. They're wetting themselves because their data shows that the ABS optioned bikes have lower claims... there might be something in that, but I suspect what it shows is that conservative riders have ABS. So HLDI are saying every bike should have ABS so they can realise a massive saying across the insurance portfolio... what do they think will happen when every bike has ABS? It wont be a 22% saving across the board that's for sure - they just normalised the population and probably removed the ABS selection bias. Conservative riders should protest right away!

  14. Keep in mind that on a less expensive bike there MIGHT be some shortcuts in the quality of the ABS. The benefit you're alluding too, might not be what you think it could be.


    Most Aussie riders have one bike and they'll commute on that bike. Whether it was chosen as a commuter primarily or recreational purpose primarily is worth thinking about. There's no two ways about it, part of the explosion in motorcycling is to do with commuting - in the scooters and smaller bikes are popular commuters.

    What benefit will the commuting market realise from ABS? The rider who rides in all conditions might realise some "benefit". I agree with you that if the kind of riding that's done is mostly fine conditions in predictable circumstances, the ABS may never activate.

    You've pointed out another confound - the studies have focused on bikes with a strong overlap in their intended use - open ride riding. They are drawing conclusions for all bikes and bike types from this subset. Interesting.
  15. I know it's colloquial, but the times where I've felt the ABS kick in and I think have stopped me sailing into a car or something like that, haven't been when I was at any terrible risk of death. It's usually when I've been commuting, mixed up in traffic, doing like 30 or 40 and someone dives into my lane and I need to jump on the brakes, and the wet weather + road surface has caused the brakes to lock up, and ABS has saved me from dropping it.

    My personal experience with ABS on my machine since November 2010 has been very positive. I don't think it's made me "rely" on it, or forgo proper braking technique (oh, she'll be right, I have ABS!), but it's great when things are marginal and you make an error, which is easy to do.

    I have a hunch that in a lot of fatalities, perfect braking technique, or ABS, wouldn't have made a difference.

    Rider aids like ABS and TC I think are additive, not replacements, for training and technique. They add a bit of extra buffer, which in road situations, is always nice to have.

    I wish I had TC on the Sprint, maybe wouldn't have binned it then :)
    • Like Like x 2
  16. Once again, I would emphasise my position that I do not regard ABS as any kind of panacea. I consider it a useful tool and a potential benefit in the way that brakes that actually worked or tyres that actually gripped once were. I am utterly opposed to compulsion, however, because I think it should only be fitted where it has a reasonable chance of doing some good.

    If I was buying a bike principally for open road use, I wouldn't want ABS because I don't want the extra weight, complication and component count of a system that I will almost certainly never use. OTOH, buying a commuter bike like the DR, if a decent ABS system were available as an option and I had the cash available, I'd have it because a good proportion of my riding will be in circumstances where, when push comes to shove, a proper ABS system will outperform me. Horses for courses.

    That said, I've commuted in Perth for most of the last decade and had ABS for only two years of that. I've not dropped it due to a braking error, hit anything because I couldn't stop in time or had the ABS activate without deliberately provoking it so maybe I'm overestimating the risk.
    • Like Like x 1
  17. Here's a take on the NHTSA's report from another forum - I'm still chewing over this take:

  18. statistics make my brain hurt
  19. But only 72% of the time \\:D/

    Cheers Spocky
    • Like Like x 4
  20.  Top